Open MohieEldinMuhammad opened 1 year ago
Same Here!
Even I get very bad results on the training images that I had taken with 100% accuracy Is it the same for you or do you get good accuracy on the training data in the test phase? @MohieEldinMuhammad @amroghoneim
@amroghoneim @ftmasadi @MohieEldinMuhammad Any updates guys? I have a similar problem.
Even I get very bad results on the training images that I had taken with 100% accuracy Is it the same for you or do you get good accuracy on the training data in the test phase? @MohieEldinMuhammad @amroghoneim
actually i was overfitting on 2 images only to test the whole pipeline first, so the problem now is that the fine-tuned model is not working properly on deployment phase, hence no need to wast time in training unless the problem is solved
@amroghoneim @ftmasadi @MohieEldinMuhammad Any updates guys? I have a similar problem.
No, unfortunately.
@amroghoneim @ftmasadi @MohieEldinMuhammad Any updates guys? I have a similar problem.
May I ask what exactly is your problem? And in which part of your results are there problems? @Arunavameister
Even I get very bad results on the training images that I had taken with 100% accuracy Is it the same for you or do you get good accuracy on the training data in the test phase? @MohieEldinMuhammad @amroghoneim
actually i was overfitting on 2 images only to test the whole pipeline first, so the problem now is that the fine-tuned model is not working properly on deployment phase, hence no need to wast time in training unless the problem is solved
Thank you for your reply But I still don't understand what I should do to solve my problem, even though I spent a lot of time on it. Is it possible for you to explain more about this problem? Thank you very much for your help @MohieEldinMuhammad
@ftmasadi I am getting random results during test predictions even though in training I had about 75% accuracy. I will post an update if I manage to make it work
@Arunavameister thanks for the help, i hope you will figure it out 🙏
Thank you very much for your reply. What language do you work on? Because I think this problem is mostly in languages that have a continuous structure like Arabic, Farsi, and Urdu because I didn't encounter such a problem when I used the English database of the site. What is your opinion about this? @Arunavameister
Facing same issue . I am trying to solve this problem. I'll report here if I handled this problem.
thanks for the help, I still don't understand what I should do to solve this problem @masoudMZB
Facing same issue. Any updates on this?
@hayderkharrufa @ftmasadi there are many reasons why you can't get proper ouput
Do you guys have any updates? Im also getting 85% accuracy when training. And then when I try to use the model with THE SAME pictures as I used for training/evaluation I get much lower accuracy.
Following this link instructions https://github.com/JaidedAI/EasyOCR/blob/master/custom_model.md. I have trained a custom model on my own dataset.
Here is the .yml file I used:
The .py file:
Then i predict like this: